AI21 is the latest company to release a competitor to GPT-3 — currently the gold standard of supercomputer-powered, AI auto-text generators.
Highly sophisticated auto-text writers like AI21 and GPT-3 easily auto-generate content such as emails, ad copy, one paragraph blog posts, poetry and the like with just a few lines of text input from a user.
Dubbed Jurassic-1, AI21’s alternative to GPT-3 weighs in at 178 billion parameters and is available in a limited free trial to everyone.
It’s also billed as the largest English language AI auto-text generator on the market.
This article in VentureBeat offers the results of a test-drive of AI21’s auto-text generator.
VentureBeat’s verdict: Jurassic-1 is a promising alternative to GPT-3, with some limitations, according VentureBeat writer Abhishek Iyer.
Once the only game in town in terms of highly sophisticated auto-text generation powered by a supercomputer, GTP-3 has been seeing new competitors challenge its crown in 2021.
Earlier this year, China released Wu Dao 2.0 — a supercomputer-powered, AI auto-text generator ten times more powerful than GPT-3 — according to its creators.
And a group of open source AI enthusiasts released Eleuther to the public this year — a free, somewhat less powerful alternative to GPT-3, powered by a network of computers.
(For an in-depth look at GPT-3, check out: “GPT-3 and AI Writing: Stunning, if Imperfect,” by Joe Dysart.)
In other AI-generated writing news:
*In-Depth Review: Kafkai: Review site MakeMoneyBro.com makes no bones about its thumbs-down review on auto-text generator Kafkai.
Observes MakeMoneyBro.com: Kafkai “promises so much yet delivers so little.
“Never in a million years would I recommend this service in its current state.”
KafKai’s Achilles Heel: The auto-text generator markets itself as a creator of long-form articles averaging 500-900 words.
The problem: Long-form articles so far have been tough to pull-off for auto-text generators powered by predictive analytics language engines like GPT-3.
Adds MakeMoneyBro.com: “Kafkai frankly generates pretty much junk content.”
*New Grammarly — Just for Computer Coders — Released: Grammarly — a popular AI writing assistant that helps users correct and stylize their writing — has released a version of its software just for developers.
Dubbed ‘Grammarly for Developers,’ the software offers coders real-time writing coaching as they develop new applications.
Specific feedback offered by Grammarly for Developers includes:
~Correctness (grammar and writing mechanics)
~Clarity (conciseness and readability)
~Engagement (vocabulary and variety)
~Delivery (tone, formality, and confidence).
*AI-Powered DeepL Translation Says it Has a Better Way: Free Web language translator DeepL says its system is outdoing much larger competitors.
Some critics agree.
Observes Anna Wyndham, a writer for Slator: “Journalists across Europe agreed.
“They said DeepL was ‘well disposed to grasping the sense of the sentence rather than being led astray by a literal translation.”
Other benefits cited by critics: Deepl produced more ‘French-sounding’ expressions.
And the tool has outdone all the tech giants — raising the bar for machine translation, according to Wyndham.
So far, more than a billion people have used DeepL, which offers translation in 20+ languages, according to Wyndham.
*Language Analysis Tool Snares New Funding: Cortical — an AI analysis software that classifies and extracts data from unstructured documents — has snared $6 million in new funding.
The tool uses neuroscience and ‘semantic folding’ to unearth the meaning of the unstructured sentences its processing — going beyond simple keyword search.
Moving forward, Cortical plans to fine-tune what it calls semantic supercomputing — the ability to process streams of natural language content at massive scale in real-time through the use of hardware acceleration.
*AI and Journalism: The Take From The University of Edinburgh: Researchers led by the University of Edinburgh have released an in-depth report on the current state of AI in journalism.
Its key findings:
~News organizations are using AI for newsgathering, production and distribution, but relevant skills and technology are unequally distributed across the industry.
~Machine learning presents opportunities to augment journalism by automating routine tasks and enabling greater scale, speed and efficiency.
~Machine learning also has the potential to help improve the depth, diversity and appeal of news.
~Concerns have been raised about legal, ethical and professional implications of its use in the newsroom due to issues of bias, ‘black box’ systems and value alignment.
~Assessments of the impacts and implications of AI on journalism to date are limited.
*AI and Journalism: Perspective from Two AI Evangelists: Two heavy promoters of AI in journalism — Ard Boer and Cecilia Campbell — offer their take on the tech’s impact in this 40-minute podcast.
Boer is product owner at NDC Mediagroep.
Campbell is chief marketing officer at United Robots, a pioneer in AI automated text generation that has had great success — mostly in Scandinavia.
Some highlights:
~Says Boer: “The most difficult part about getting your news organization to use automated content is explaining the concept to your colleagues.
“Most journalists, they’re not that tech savvy – at least not for a regional publisher like us.”
~Says Campbell: “There’s a Web publisher in Norway called Bergens Tidende, and they have built a whole vertical on their site about home sales.
“It’s entirely populated with the automatically generated content.
“With the readers and the ad revenue that they actually gain from that vertical — it far outperforms the cost of the robot content.”
*JournalismAI Gears for Online Conference, Nov. 29 – Dec. 3: Journalism think-tank Polis is putting together an online conference emphasizing the positive impact of AI in journalism.
Dubbed the JournalismAI Festival, the conference will essentially be a free crash course on AI and journalism — serving up talks and presentations from numerous experts and practitioners of AI-driven journalism.
You can get a feel for what the conference will be like by checking out videos recorded at the group’s conference from last year.
Underwritten by the Google News Initiative, the fest offers editors, writers, AI experts and others an up-to-the moment look at how AI is changing the news business — delivered from the perspective of the companies and news publishers championing that change.
*Some Sqawk at Stanford’s New AI Center: Some AI researchers have problems with a newly founded AI think tank at Stanford University — the Center for Research on Foundation Models.
The problem: The researchers say Stanford’s decision to spotlight supercomputer-driven AI auto text generators like GPT-3 as ‘foundational’ is off-the-mark.
Says Jitendra Malik, an AI professor at UC Berkeley: “I think the term ‘foundation’ is horribly wrong.
“These models (like GPT-3) are really castles-in-the-air — they have no foundation whatsoever.
“The language we have in these models is not grounded, there is this fakeness, there is no real understanding.”
*AI Big Picture: Top Four Dangers of AI: AI expert Kai-Fu Lee warns there are some dire threats to the world’s future hidden in the siren call of AI.
Says Lee: “The single largest danger is autonomous weapons.
“That’s when AI can be trained to kill — and more specifically trained to assassinate.”
Other threats Lee sees include:
~Unintended consequences
~Threats to personal data
~AI’s inability to explain consequential choices
Lee scored big as an author a few years back with his AI bestseller “AI Superpowers: China, Silicon Valley, and the New World Order.”
He has a new book out, “AI 2041: Ten Visions of Our Future.”
The tome is a collection of fictional short stories about our AI future — extrapolated from what Lee knows about the current state of AI.
Lee co-wrote the book with science fiction writer Chen Qiufan.
Share a Link: Please consider sharing a link to https://RobotWritersAI.com from your blog, social media post, publication or emails. More links leading to RobotWritersAI.com helps everyone interested in AI-generated writing.
–Joe Dysart is editor of RobotWritersAI.com and a tech journalist with 20+ years experience. His work has appeared in 150+ publications, including The New York Times and the Financial Times of London.